I have to say that I disagree with Dr Geddes. To me speakers with the widest dispersion with the best FR off axis sound best.
I alluded earlier to their being something of a divide on the program the speaker is designed to play.
Pretty much all classical music is recorded in an ambient field. In addition it is listened to live in spaces where the the bulk of the sound reaching the audience is reflected and not direct sound. However the music tends to be recorded at a place where the direct content is high. This I believe is due to speaker deficiencies and not microphone deficiencies. I have made hundreds of recordings. I have found that my speakers and some others will reproduce music recorded in the far field, but the vast majority will not.
So it means that for a speaker to reproduce this type of program accurately then the listener needs as much reflected sound as possible. However this only works if the speaker has a pretty accurate FR response in the off axis response.
Pop music on the other hand is recorded usually with the ambient filed excluded as much as possible and ambiance added back artificially. In this situation I suspect some control and narrowing of dispersion may be beneficial.
I'm not sure about this though. I don't like or listen to pop and rock, and make no attempt to design for it in my systems. However individuals seem to want to come here to check mixes and like the sound. So I remain uncertain as to whether speakers do have to be designed with a specific program in mind. I do know that speakers with an "up front" sound are hopeless for the reproduction of classical music.
This is the dispersion and FR of my mains.
The black line is 90 degrees off axis, the blue line is on axis, the others points between.
As you know these speakers are large, and you can decide if they are ugly. However they are truly full range and high spl. They produce a very deep depth of field and the sound is not up and close and in your face. There is plenty of detail.
I have to say my favorite tweeters are the best soft dome tweeters.
These speakers use
this one.
Concerning the topic at hand the most important observation I would make is that you can not judge a tweeter in isolation. You are inevitably judging it based on the performance of the driver below it. In fact I would suggest you are more often than not you are judging the driver below the crossover to the tweeter than the tweeter.
As you know I continually stress the importance of selecting wide band drivers. The problem again is dispersion. Getting good driver match to a tweeter is important and difficult. If crossover can truly be achieved before break up of the mid or mid woofer, then it makes life much easier. If motion is still pistonic the dispersion will be excellent, and it is easier to mate to a high dispersion tweeter. Most drivers are breaking up before crossover. It is here most problems arise, as only expensive drivers are not going to break up by a good crossover point to a tweeter. Getting the crossover either out of the speech discrimination band or pushing it to the top end of the speech discrimination band I believe has great benefit.
The other option is to lower the crossover point. This however increases the power drive to the tweeter and risks burn out.
In addition unless the tweeter is down 24 db electrically by resonance it will sound rough, unless it is a very low Qts driver. Those are not common.
In the speaker I illustrated the light rigid cones are pistonic to crossover at 2.5 KHz, but break up violently less than an octave above crossover which has to be dealt with. Another half octave before break up would be advantageous. However as always this is where compromise comes in. You just have to cut your cloth to the units you have selected.
If you are mating a tweeter to a drive unit already in irregular break up then more than likely you will blame the tweeter.
Lastly Klipsch no longer use horn loaded compression drivers. They now use soft dome tweeters with a horn in front.
In addition they are crossing over at 1500 Hz. They pay the penalty with a higher incidence of tweeter VC burn out. I have no idea of the Fs or Qts of those tweeters.
